Transmission of directly modulated lasers (or other light sources) through optical components results in generation of second-order distortion due to the chromatic dispersion properties of the optical components and chirp (i.e., wavelength variations) of the laser (or other light source). Within this application the term optical components is used to denote optical devices that exhibits chromatic dispersion, meaning that the time delay of the signal through the device varies with the optical wavelength. This includes, but is not limited to, both passive components such as optical fiber, optical filters, multiplexers, waveguides, etc. and also active optical components such as optical amplifiers (e.g., EDFAs, Raman amplifiers), modulators, integrated devices, etc.
The composite second-order (CSO) distortion is an indication of the severity of these second-order distortions. CSO distortion due to the chromatic dispersion properties of optical components degrade the performance of continuous-wave (CW) analog signals and pseudo-analog signals such as quadrature-amplitude modulated (QAM) signals.
FIG. 1A illustrates a fiber-optic system including a conventional directly modulated laser whose optical output goes through an optical component and is then detected by the optical receiver Rx. The optical output of the laser is directly modulated by a current signal I(t) which contains the signal information. The light from the laser goes through the optical component, and the optical output of the component is denoted by POUT(t). The photodiode current in the receiver is denoted by IPD(t).
The optical component is treated as a linear transmission element that is described by two parameters: optical attenuation a and delay T. The chromatic dispersion of the optical component results in the delay being a function of wavelength; that is, τ=τ·(λ). The combined effects of chromatic dispersion and laser chirp is to delay different parts of the input waveforms by different amounts, which results in an amplitude correction factor of
      [          1      +                        ∂          τ                          ∂          t                      ]        -    1  that causes distortion of the input optical waveform. Consequently, the output of the optical component is given by the equation:
                                          P            OUT                    ⁡                      (            t            )                          =                              α            ⁢                                                  ⁢                                          P                IN                            ⁡                              (                                  t                  -                                      τ                    ⁡                                          (                      λ                      )                                                                      )                                                          [                          1              +                                                ∂                  τ                                                  ∂                  t                                                      ]                                              (                  Equation          ⁢                                          ⁢          1                )            as shown in FIG. 1A.Dispersion Compensation Techniques
Several strategies are known to those skilled in the art to avoid or suppress such second-order distortions. One is to tune the wavelength of the laser to the zero-dispersion wavelength of the optical component. The CSO distortion that is generated by the dispersion of the optical component then becomes negligible.
However, sometimes it is not possible to tune the laser wavelength to the zero-dispersion wavelength of the optical component. For example, if the laser transmitters are used in a DWDM system, then the laser wavelengths are constrained to specific values defined by standard groups such as the ITU and cannot be arbitrarily tuned. An example is a DWDM system operating in the C-Band (1550 nm region) of conventional single-mode fiber, the zero-dispersion wavelength is near 1310 nm and therefore the laser wavelengths cannot be tuned to the zero-dispersion wavelength of the optical fiber. In this case, the dispersion of the optical component (the optical fiber in this instance) is unavoidably high.
In those systems where the dispersion of the optical component is high, other techniques to suppress second-order distortions are known to those skilled in the art. There are several inventions that describe laser predistortion circuits to compensate for the second-order distortions generated by dispersion.
Dispersion-Slope Compensation Techniques
There are cases where the optical components also have extremely high values of dispersion-slope. “Dispersion-slope” refers to the slope of the dispersion characteristic of the optical components when plotted as a function of the optical wavelength. For example, dispersion slopes as high as 150 ps/nm2 have been measured in optical filters used in DWDM systems. In comparison, 20 km of conventional single-mode fiber has a dispersion slope at the zero-dispersion wavelength of less than 2 ps/nm2.
That is, the dispersion-slope of some optical components can be as high as that of 1600 km of optical fiber. At these high values of dispersion slopes, second-order effects (modulation of fiber delay due to residual dispersion that arises from the second-order term in the power series expansion of the delay characteristic) generates significant CSO distortion that can degrade the performance of CATV systems that employ analog signals or quasi-analog signals such as QAM.
It has been found that the dispersion-slope-induced CSO distortion can be as severe in magnitude as the CSO generated by the dispersion of the optical components at high dispersion-slope values. That is, second-order distortions can be generated both by the dispersion and dispersion-slope of optical components. This disclosure is concerned with suppressing the distortions generated by the dispersion-slope of optical components.
Prior inventions that describe suppression of second-order distortions generated by the dispersion of the optical components do not apply to the suppression of CSO generated by the dispersion-slope of the optical components. There are other inventions, however, that deal with suppression of CSO distortions generated by the dispersion-slope of optical components. They are generally referred to as dispersion-slope compensation techniques.
These prior dispersion-slope compensation inventions generally involve adding another device (or fiber) after the laser transmitter that has a dispersion-slope opposite in sign and equal in magnitude to the dispersion-slope of the optical component whose distortions you are trying to suppress. These other inventions suffer from two major disadvantages compared to the instant disclosure: (1) They usually compensate for only one value of dispersion-slope. Therefore, the dispersion-slope compensation device has to be tailor-made for the optical component whose distortions you are trying to suppress. If the optical component is replaced then the compensation device also has to be replaced. (2) The dispersion-slope compensation device is usually expensive—often as expensive, or more expensive, than the optical component whose distortion one is trying to suppress.
The present disclosure overcomes these limitations of previous inventions by performing dispersion-slope compensation using electronic compensation circuitry rather than using optical devices that have a dispersion slope opposite in sign to the component whose dispersion-slope one is trying to compensate. The compensation circuitry is also sometimes referred to as predistortion circuitry, but this must not be construed as limiting the location of this circuitry to a point prior to the optical source, such as in the optical transmitter. Since the generation of CSO distortion due to the dispersion and dispersion-slope of optical components is a linear process, the compensation circuitry can be placed anywhere in the optical network—either in the optical transmitter prior to the light source or in the optical receiver after the optical photodiode.
The cost of the electronic compensation circuitry describe in this disclosure is negligible since the technique adds a few inexpensive electronic parts to the existing optical transmitter or receiver. Furthermore, the compensation circuitry of this disclosure can compensate for the distortion generated by any value of dispersion-slope, and could even be made adaptive so that the technique remains effective even if the dispersion-slope of the optical component changes over time—and is therefore superior to prior inventions.